Basecamp Cloud Fundamentals

Cloud Computing for People Who've Never Touched a Server

IaaS, PaaS, and SaaS explained through renting, not buying — and where to get a free server to practise on.

"The cloud" gets talked about like it's a weather system — vague, everywhere, slightly mystical. It isn't. The cloud is other people's computers, sitting in enormous warehouses called data centres, that you pay to use over the internet instead of buying and running the equivalent hardware yourself.

That's genuinely most of it. The rest of this post is just working out what "use" means, because that word covers a surprisingly wide range of arrangements.

Renting a car vs. renting a driver

The clearest way to understand the different flavours of cloud service is to think about how you'd get somewhere without owning a car.

Cloud computing maps onto exactly this spectrum:

IaaS — Infrastructure as a Service (renting the car)

You get raw virtual machines, storage, and networking. You still install the operating system, patch it, configure it, and manage it — but you never had to buy a physical server, rack it, or replace a failed hard drive. Examples: Azure Virtual Machines, AWS EC2.

PaaS — Platform as a Service (car with a driver)

You hand over your application code, and the provider handles the operating system, patching, and scaling underneath it. You focus on your code; they focus on keeping the platform running. Examples: Azure App Service, AWS Elastic Beanstalk.

SaaS — Software as a Service (the taxi app)

You just use the finished product through a browser. No infrastructure to think about at all. Examples: Gmail, Microsoft 365, Slack — most of the software you already use daily without calling it "cloud" in your head.

Trail note

Notice the pattern: as you move from IaaS → PaaS → SaaS, you get less control but less responsibility. That trade-off — control versus responsibility — is the single most useful lens for understanding almost any cloud architecture decision you'll encounter later.

Why companies actually moved to it

Before the cloud existed at scale, a company that suddenly needed more computing power had to buy physical servers, wait for delivery, rack them, and configure them — a process measured in weeks. If that surge in demand turned out to be temporary, they were now stuck owning expensive hardware sitting mostly idle.

Cloud providers turned that into something you could switch on in minutes and switch off just as fast, paying only for what you actually used. That single shift — from buying capacity you hope you'll need, to renting capacity the moment you need it — is the entire economic argument for cloud computing. Everything else is a variation on that theme.

What "the cloud" physically is

It's worth grounding this in something concrete: a cloud provider's data centre is a real building, often the size of several warehouses, packed with server racks, cooling systems, and redundant power supplies. When you spin up a virtual machine on a public cloud, you are — however indirectly — renting a slice of physical hardware sitting in one of those specific buildings, in a specific region of the world you usually get to choose. It's not abstract. It's just far away and shared.

Get your hands on it, for free

You don't need permission from an employer to start learning this hands-on:

The vocabulary will keep coming — regions, availability zones, resource groups, subscriptions — but none of it will stick properly until you've actually clicked "create" on something and watched it come to life.

Next waypoint — Trailhead

Your First 30 Days in Cybersecurity: What Actually Matters →